Brand-Category Exposure Panel · Rubric v1.0.2

nvidia.com × Data Center GPUs

Week 16 2026 · nvidia.com · Source: raw_mentions
Computed 2026-04-24 17:34Z
Composite exposure
0.46
/ 1.00
medium
Five-factor composite over 5 contributing factors. Each factor independently scored, weighted per Rubric v1.0.2, and composited. Confidence reflects the weakest contributing factor.
Factor contributions
AI attention
Geographic concentration
Chokepoint exposure
Vendor concentration
LLM consensus
AI attention
How concentrated 13 LLMs are on this category for this brand
high
1.000
/ 1.000
At ceiling. Nvidia × DCG is the most concentrated brand-category cell in the week.
Weight 0.25Contribution 0.250 (55%)
Geographic concentration
How clustered the facility set is across 500km geographic bins
high
0.170
/ 1.000
Distributed. Facility set spans multiple regional clusters.
Weight 0.30Contribution 0.051 (11%)
Chokepoint exposure
Fraction of relevant facilities within 200km of a curated chokepoint
medium
0.201
/ 1.000
Partial exposure. A meaningful fraction of facilities in the radius.
Weight 0.20Contribution 0.040 (9%)
Vendor concentration
Qualified-operator pool tightness in the leading-edge cohort
high
0.750
/ 1.000
Stressed environment. 3 qualified operators in the cohort.
Weight 0.15Contribution 0.113 (25%)
LLM consensus
Inter-provider agreement on category attribution (inverted to risk)
medium
0.030
/ 1.000
Strong consensus. Providers agree closely on category attribution.
Weight 0.10Contribution 0.003 (1%)
Interpretation

The nvidia.com × Data Center GPUs supply chain is not Taiwan-monolithic. Across the ~200 relevant facilities spanning the leading-edge foundry cohort (TSMC / Samsung Foundry / Intel Foundry), the HBM memory cohort (SK Hynix / Samsung Semiconductor / Micron), and the advanced-packaging cohort (TSMC / ASE / Amkor), only the Taiwan cluster sits within 200km of the chokepoint in Rubric v1.0.2's seed.

Geographic concentration (F2) and chokepoint exposure (F3) both land lower than the naive illustrative in the rubric's worked example, because the actual supply chain spans 4 regional clusters rather than clustering in Taiwan. The higher-weighted attention and consensus factors (F1, F5) pull the composite toward medium exposure.

Rubric v1.0.2 · spec · Provider mapping v1.0
Factors present: 5/5